Article
pubs.acs.org/JPCC
Catalytic Mechanisms of Sulfur-Doped Graphene as Efficient Oxygen
Reduction Reaction Catalysts for Fuel Cells
Lipeng Zhang, Jianbing Niu, Mingtao Li, and Zhenhai Xia*
Department of Materials Science and Engineering, Department of Chemistry, University of North Texas, Denton, Texas 76203,
United States
ABSTRACT: Density functional theory (DFT) was applied
to study sulfur-doped graphene clusters as oxygen reduction
reaction (ORR) cathode catalysts for fuel cells. Several sulfurdoped graphene clusters with/without Stone−Wales defects
were investigated and their electronic structures, reaction free
energy, transition states, and energy barriers were calculated to
predict their catalytic properties. The results show that sulfur
atoms could be adsorbed on the graphene surface, substitute
carbon atoms at the graphene edges in the form of sulfur/
sulfur oxide, or connect two graphene sheets by forming a
sulfur cluster ring. These sulfur-doped graphene clusters with
sulfur or sulfur oxide locating at graphene edges show
electrocatalytic activity for ORR. Catalytic active sites distribute at the zigzag edge or the neighboring carbon atoms of
doped sulfur oxide atoms, which possess large spin or charge density. For those being the active catalytic sites, sulfur atoms with
the highest charge density take a two-electron transfer pathway while the carbon atoms with high spin or charge density follow a
four-electron transfer pathway. It was predicted from the reaction energy barriers that the sulfur-doped graphene could show
ORR catalytic properties comparable to platinum. The prediction is consistent with the experimental results on S-doped
graphene.
and mesoporous graphitic arrays,26 exhibit high electrocatalytic
activity and CO tolerance in comparison to conventional
platinum catalysts for ORR and are promising candidates for
replacing Pt-based catalysts. B-doped27 and N/B-codoped
graphene28 also show high catalytic property for ORR. The
high activity of these doped graphenes (nitrogen, nitrogen/
boron) may be attributed to the polarized distribution of spin
and charge density29,30 which are caused by the introduced
heteroatoms. More recently, sulfur-doped graphene has been
synthesized by using different methods,12,31,32 which exhibits
competitive catalytic activities compared to nitrogen-doped
graphene and even better catalytic activities than commercial
Pt/C. Experimental results show that the onset potential and
the number of transferred electrons per oxygen for S-doped
graphene are close to those for the nitrogen-doped
graphene.12,19 The sulfur-doped graphene expands the family
of metal-free carbon-based nanomaterials as a new electrocatalyst to replace Pt in fuel cells.
Density functional theory (DFT) is an effective theoretical
method to study the electronic property of catalytic materials
and ORR pathways. The prominent pathway of ORR in proton
exchange membrane (PEM) fuel cell and the kinetics of the
proposed nonelectrochemical reactions were studied by using
the DFT method.33 The mechanisms of ORR on carbon-
1. INTRODUCTION
High conversion efficiency, high power density, quiet operation,
and no pollution are the remarkable advantages of fuel cells for
various applications. However, the kinetics of the oxygen
reduction reaction (ORR) on cathode is sluggish without
catalysts.1,2 In principle, the ORR can process through direct
four-electron transfer pathway, O2 + 4H+ + 4e− → 2H2O, or
two-electron transfer pathway in which hydrogen peroxide
formed, O2 + 2H+ + 2e− → H2O2. The former pathway is
expected to occur to achieve high efficiency. Thus, a route to
search for an efficient catalyst is to determine if this catalyst
facilitates the four-electron pathway. So far, the most effective
electrocatalyst for ORRs on cathode is Pt or its alloys, which
proceeds through four-electron transfer.3,4 However, the high
cost, limited supply, poor durability, and stability of Pt have
hindered the large-scale application of fuel cells. Therefore, the
search for new nonprecious metal5−8 or metal-free9−12 catalysts
with high activity and practical durability has received a great
deal of interest.
Graphene, a two-dimensional monolayer structure of sp2
hybridization carbon, has attracted great attention in a wide
range of fields, such as electronics,13,14 sensors,15,16 batteries,17,18 and catalysts,19 due to its exceptional properties.20−22 Both theoretical and experimental studies have
revealed that doped heteroatoms such as nitrogen and boron
can modify their electrical properties and chemical activities.23
Recently, studies have confirmed that N-doped carbon
materials, such as carbon nanotubes (CNTs),9,24 graphene,19,25
© 2014 American Chemical Society
Received: October 23, 2013
Revised: January 3, 2014
Published: January 29, 2014
3545
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
Figure 1. Several possible sulfur-doped graphene clusters: (a) sulfur atoms adsorbed on the surface of the graphene cluster; substituting sulfur atoms
at (b) zigzag and (c) armchair edges; SO2 substituted at (d) zigzag and (e) armchair edges; and (f) sulfur ring cluster connecting two pieces of
graphene. The structures of the graphene are shown only partially to highlight the doping structures. Small white, gray, yellow, and red balls
represent hydrogen, carbon, sulfur, and oxygen atoms, respectively.
The former two peaks were corresponding to 2p3/2 and 2p1/2
positions of thiophene-S due to their spin−orbit coupling. The
third peak related to some oxidized sulfur. Binding energies
around 162.0 eV (S−H) and higher than 165.5 eV were not
found. So the sulfur are inferred to be mainly doped at the
edges or on the surface of graphene in the form of −C−S−C−
or −C−SO2−C−. To reduce the calculation expense and show
the effect of sulfur bonding structures on the local electronic
properties of graphene at the same time, we model a sheet of
graphene with 100 carbon atoms, the edged carbon atoms of
which are terminated by hydrogen atoms. These graphene
clusters are large enough to study the local effect of doping
while having good computational efficiency. In these doped
graphene clusters, Stone−Wales defects were also introduced to
study defect effects. Stone−Wales defects are one type of
important topological defects in sp2-bonded carbon materials,
which could affect the electronic property of graphene. The
optimization structures of all these sulfur-doped graphene
clusters were calculated by using the DFT. Formation energies
of these SGC were calculated as follows: Ef = ES‑graphene + yμC −
(Egraphene + xμS/S‑oxide), where ES‑graphene is the energy of SGC,
Egraphene is the energy of the corresponding graphene cluster, μC
is the chemical potential of C, and μS/S‑oxide is the chemical
potential of S8 or sulfur oxide (SO2), respectively.
The ORR processes were simulated to explore possible
reaction pathways in the presence of SGC. In an acidic
environment, a unified mechanism for the first reduction step,
which combines Damjanovic’s proton participation in the first
electron reduction step and Yeager’s dissociative chemisorptions of O2, is summarized as follows as Path I
supported Fe-phthalocyanine (FePc/C) and Co-ptthalocyanine
(CoPc/C) in alkaline solution were also elucidated.34 Anderson
et al. applied the DFT method on studying the oxygen
reduction on graphene, nitrogen-doped graphene, and cobalt−
graphene−nitride systems.35−37 Recently, Zhang et al. using the
DFT method studied the ORR mechanisms on the nitrogendoped graphene29 effect of the microstructure (the number of
dopants and defects) of nitrogen-doped graphene for ORR in
acidic environment.30 They proposed a four-electron ORR
pathway on N-doped graphene, and that Stone−Wale defects
facilitate the ORR on nitrogen-doped graphene. Furthermore,
DFT calculations were also used to study the mechanisms of
ORR on N-doped graphene in an alkaline environment by Yu
et al.,38 who took the solvent, surface coverage, and adsorbates
into consideration, and obtained the overall energy profile of
the ORR pathway. Although the nitrogen/boron-doped
graphenes have been studied theoretically,39−41 little work
was done on the sulfur-doped graphene, and their effect on the
ORR. Sulfur belongs to the p-block of the periodic table, which
have unique electronic structures (p orbitals in the outermost
shell) similar to nitrogen, but with different electronegativities.
It is of interest to explore the catalytic mechanism of S-doped
graphene. In this work, using DFT calculation, we studied the
doping structure of sulfur atoms on the graphene clusters and
their catalytic mechanism for ORR in an acidic environment.
These sulfur-doped graphene clusters show electrocatalytic
activity for ORR, which strongly depends on their doping
structures.
2. METHODS
B3LYP hybrid density functional theory (DFT) of Gaussian 09
(Revision A. 02; Gaussian, Inc.: Wallingford, CT, 2009) was
employed with a basis set of 6-31G (d,p).29 Four possible types
of sulfur-doped graphene clusters (SGC) were considered, as
schematically shown in Figure 1. Type one is sulfur atoms
adsorbed on the surface of the graphene cluster (C100H26S).
Type two represents the sulfur atom substitution at the zigzag
or armchair edge of the graphene cluster (C99H25S). In Type
three, the sulfur atom substitutes the carbon atoms at the
graphene edge (zigzag and armchair) in the form of −C−SO2−
C− (C99H25SO2). The last type is two pieces of graphene
clusters connected by a sulfur ring (C 144 H 40 S 4 ). For
comparison, pure graphene (C100H26) was also analyzed in
this study. These models of sulfur bonding structures are built
based on the experimental structure analyses of sulfur-doped
graphene.31,42,43 X-ray photoelectron spectroscopy (XPS)
shows all the high-resolution S2p peaks of sulfur-doped
graphene could be resolved into three different peaks at
binding energies of ∼163.9, 165.1, and 168.9 eV, respectively.
O2 + H+ → OOH+
(1)
OOH+ + * + e− → *OOH
(2)
or Path II
O2 + * + e− → *O−O−
(3)
*O−O− + H+ → *OOH
(4)
where the asterisk represents a chemisorption site on the
graphene cluster. In Path I, O2 first reacts with a proton to form
OOH+30 and then adsorb on active sites of the graphene cluster
after the first electron transmission was completed. To examine
this reaction path, we set an OOH molecule near the graphene
cluster plane at a distance of 1.5−3 Å, and then observed if it
adsorbs on the graphene surface. We also set the O2 molecule
near the graphene to check whether it could adsorb at the
potential catalytic active sites or not. After the first electron
transformation, the succeeding electron transforming was
3546
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
Table 1. Formation Energy (eV) of S-Doped Graphene Clusters
graphene
clusters
without
defect
with defect
surface
adsorption P 1
surface
adsorption P 2
zigzag edge
substitution
armchair edge
substitution
SO2- zigzag edge
substitution
SO2- armchair edge
substitution
ring
clustering
−1.46
−1.73
0.90
2.02
1.47
2.73
2.70
−2.60
−2.31
0.89
1.80
1.44
2.37
2.54
Figure 2. Atomic charge density and spin density distributions on the S-doped and pure graphene clusters. Atomic charge density distribution on Sadsorbed graphene clusters with (a) perfect structure and (b) one Stone−Wales defect; (c) atomic charge density and (d) spin density distributions
on perfect graphene cluster with substituting S at the zigzag edge; (e) atomic charge density and (f) spin density on SO2-doped graphene with a
Stone−Wales defect; atomic charge density on (g) sulfur ring cluster connecting two pieces of graphene clusters and (h) pure graphene cluster. The
colors of the balls stand for relative values of charge and spin density. The density decreases linearly from positive to negative values in the color
order of red, orange, yellow, green, and blue. Sulfur and oxygen atoms are labeled with S and O, respectively. The unlabeled small and large balls
represent H and C, respectively.
3. RESULTS AND DISCUSSION
3.1. Doped Graphene Clusters and Formation Energy.
Figure 1 shows four possible sulfur-doping graphene clusters,
namely, sulfur chemisorption on surface, S substitution at edge,
SO2 substitution at edge, and sulfur ring clusters, which have
been described in detail in the Methods section. The formation
energies of these sulfur-doped graphenes were calculated and
are listed in Table 1. The formation energies for sulfur
adsorbing on the graphene surface (Figure 1a) are negative, but
they are positive for sulfur or sulfur oxide substitution at the
edges of graphene clusters (Figure 1b−e) and the sulfur ring
cluster connecting graphene clusters (Figure1f). Therefore,
compared to the sulfur (or sulfur oxide) edge substitution
(Figure 1b−e) or sulfur ring cluster connecting graphene
(Figure 1f) sulfur adsorption on the graphene surface is
energetically favorable. In the presence of Stone−Wales defects
on the graphene cluster, the formation energies of sulfur-doped
graphene clusters are lower than those of perfect graphene
simulated by adding H atoms in the system. For each step, we
obtained the transition state structures and optimized
structures, and calculated the reaction energy barrier ΔEb and
reaction free energy ΔG. ΔEb is defined as the difference
between the energy of transition structures (ET) and initial
structures (EI), ΔEb = ET − EI, and ΔG is the difference
between free energies of the final and initial states given by the
following expression:44,45 ΔG = ΔE + ΔZPE − TΔS, where ΔE
is the reaction energy calculated by the difference of chemical
potential between product and reactant molecules adsorbed on
the catalyst surface, obtained from DFT calculations of
optimization structures, ZPE is the zero point energy, S is the
entropy, which are obtained by calculating the frequency of
optimization structure, and T is the temperature. For the
reaction with negative reaction free energy, it would occur
spontaneously.
3547
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
Figure 3. ORR process on the sulfur-doped graphene cluster when the catalytic active site is the sulfur atom: (a) two OOH molecules adsorbed on
the sulfur atom and (b) two H2O2 molecules formed and departed from the sulfur atom after the introduction of two more H atoms.
those above, the atoms with high charge density are also located
at the zigzag edge. Furthermore, spin density is also introduced
on the atom at the edge. For example, the edge carbon atoms
11 and 15 possess the highest charge density of 0.19 among the
carbon atoms. For spin density, edge atom 10, with the largest
value of 0.39, and edge atoms 6 and 14, with the second largest
value of 0.27, are found on the graphene. In addition, the sulfur
atom (number 88) obsesses a maximum positive charge density
of 0.22. Similar spin and charge density distributions can be
found on the defective graphene structures with sulfur oxide
atom at the armchair edge (Figure 2e,f). Besides, the
neighboring carbons at the doped sulfur oxide also have high
spin and charge densities. Here, again, the Stone−Wales defects
make the sulfur-doped graphene clusters polarized more atoms
with higher spin and charge density can be found on the doped
graphene cluster with Stone−Wales defects than those on the
perfect one, suggesting that the defects could generate more
catalytic active sites to ORR. For the sulfur ring cluster
connecting two pieces of graphene cluster, these atoms with
higher charge density are also at the zigzag edge or neighboring
the sulfur atoms on the graphene clusters, but the value is less
than 0.19 (Figure 2g). We have tested these atoms with the
high charge or spin density and confirmed that all these atoms
with charge density larger than 0.20 and spin density larger than
0.15 can adsorbe OOH or O2 and could be potential active sites
for ORR. To demonstrate the effect of the doped sulfur atom,
we also calculated the charge density distribution on the pure
graphene cluster. As shown in Figure 2h, there is no spin
density on the pure graphene cluster, and the maximum value
of charge density is 0.17. The overall charge distribution is
similar to that on the graphene cluster having sulfur atom
chemisorption. OOH or O2 could not adsorb on any of its
carbon atoms with higher charge density. Thus the catalytic
capability of the pure graphene is limited.
3.3. ORR Pathway on Sulfur-Doped Graphene.
3.3.1. Two-Electron Transfer Paths. Once an OOH is
adsorbed on the doped graphene surface, the next step of the
ORR could be the O−O bond break, which represents fourelectron transfer. Otherwise, the ORR is a two-electron transfer.
We have examined all the possible active sites selected on the
basis of large positive charge density and spin density, and
simulated the reaction when a proton is added to the position
near adsorbed OOH. We found that the ORR is either fourelectron or two-electron transfer depending on the doping
structures. In the case of the sulfur atom being the catalytic
active site, two OOH species can adsorb on the S atom (Figure
3a). The distance between adsorbed oxygen (OOH) and sulfur
atoms decreased to 1.8 Å from the original distance of 3.0 Å
after structural optimization. Thus, the S−O covalent bond
clusters. So Stone−Wales defects facilitate sulfur doping on the
graphene clusters. This may be attributed to the fact that
defects change the local charge distribution and crystal lattice.
For the edge substitution, the difference in the formation
energy between the perfect and the defective graphene cluster
is ∼0.36 eV, lower than that of the sulfur (sulfur oxide)
adsorbing on the graphene surface or the sulfur ring cluster
graphene. Thus, Stone−Wales defects at the center of the
graphene cluster have a weak effect on the sulfur edge
substitution. For the same graphene cluster, the formation
energy of sulfur atoms (sulfur oxide) at the zigzag edge is
always lower than that of the armchair one, suggesting that
sulfur (sulfur oxide) is preferable to substitute the carbon atoms
at the zigzag edge.
3.2. Active Catalytic Sites of the S-Doped Graphene
Clusters. We calculated spin and charge densities of each atom
of the sulfur-doped graphene clusters and determined possible
ORR catalytic active sites on these doped structures on the
basis of these spin and charge density distributions. It was
shown in our prior work29,30 that the ORR catalytic active sites
are closely related to the charge and spin density distributions.
Figure 2 shows the atomic charge and spin density distributions
on the sulfur-doped and pure graphene clusters. For the sulfuradsorbed graphene surface, the sulfur atom does not introduce
an extra unpaired electron, therefore, the graphene does not
exhibit additional spin density. The charge density, on the other
hand, redistributes on the perfect or defective (Stone−Wales
defects) S-doped graphene cluster (Figure 2a,b). Specifically,
carbon atoms designated with the numbers 11, 15, 89, and 93 at
the zigzag edge possess higher positive charge density around
0.17. These carbon atoms may be the catalytic active sites for
ORR. To test this hypothesis, we have calculated the
adsorption of OOH or O2 species on these sites by setting
them near these potential catalytic sites. The adsorption of
OOH or O2 species is the first step necessary for the graphene
to catalyze the ORR. The results show that both OOH and O2
can adsorb on these atoms located at the zigzag edge of
graphene with Stone−Wales defects but cannot on the
graphene without Stone−Wales defects. Thus, those Sadsorbed graphene could have catalytic activities depending
on the Stone−Wales defects. Here, the defects play an
important role in facilitating the ORR. Compared to the
perfect graphene cluster, the sulfur-doped graphene surface
twisted a little bit when the Stone−Wales defects were
introduced, which changed the crystal lattice and local charge
distribution on the graphene cluster.
Panels c and d of Figure 2 show the atomic charge and spin
density distributions on the graphene cluster with substitutional
sulfur atoms located at the zigzag edge, respectively. Similar to
3548
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
Figure 4. ORR processes on the sulfur-doped graphene cluster where a carbon atom at the zigzag edge acts as the catalytic active site because of the
highest spin density on it: (a) OOH adsorbed on the carbon atom, (b) rupture of the O−O bond and formation of the water molecule after an H
atom was introduced into the system, (c) formation of an OH after the second H was introduced, and (d) formation of another water molecule after
the third H was introduced into the system.
3.4. Reaction Free Energy and Energy Barriers.
Reaction free energy ΔG was calculated for each substep of
the ORR over sulfur-doped graphene clusters. For the first
electron-transfer process, ΔG was determined for two different
mechanisms (Paths I and II) since Damjanovic’s proton
participation reactions and/or Yeager’s dissociative mechanism
may occur on the same graphene. Values of ΔG for two- and
four-electron transfer pathways are listed in Table 2. For the
formed and the OOH chemically adsorbed on the S atom. As
two more H atoms were set near the oxygen atoms of two
OOH species, they adsorbed to the oxygen atoms in OOH
molecules that were bonded to the sulfur atom, respectively.
The bonds between sulfur and oxygen atoms break in the
reaction, resulting in the formation of two H2O2 molecules.
Finally, H2O2 molecules moved away from the graphene
surface. The final distance between the H2O2 and graphene is
3.5 Å (Figure 3b).
3.3.2. Four-Electron Transfer Paths. For these carbon atoms
with large spin or charge density as catalytic active sites, it was
found that the four-electron transfer usually occurs. For
example, for edge S-doped graphene, OOH was able to adsorb
to atom number 10, the carbon with the highest spin density
(Figure 4a). When a proton was introduced near the adsorbed
OOH, it resulted in rupture of the O−O bond and formation of
one water molecule while one oxygen atom still adsorbed on
the graphene alone (Figure 4b). As mentioned above, the
breakage of the O−O bond is the key step of four-electron
transfer, which defines the process being the four-electron
transfer pathway. After two protons were successively
introduced into the system, as shown in Figure 4c,d, another
water molecule formed and departed from the graphene. The
finial distance between the two water molecules and graphene
is ∼3.4 Å.
The above calculations show that for S-doped graphene
those carbons with high charge or spin density facilitate fourelectron transfer while the sulfur itself promotes two-electron
transfer. As mentioned before, two-electron transfer ORR is
inefficient. Although sulfur doping activates carbon atoms as
active sites for four-electron transfer ORR, the existence of the
sulfur dopants reduces the efficiency of the ORR. However,
among these doping structures, the −SO2− bonding structure
(Figure 1d,e) does not catalyze the two-electron transfer
reactions while it activates carbon atoms for ORR. We thus
suggest that the catalytic efficiency may be improved by
introducing −SO2− bonding structures during the graphene
doping process.
Table 2. Reaction Free Energy, ΔG (eV), of Two-Electron
and Four-Electron Transfer Reaction Processes on SulfurDoped Graphene with and without Stone−Wales Defectsa
sub-reactions
two-electron transfer
pathway
four-electron transfer
pathway
O2 + H+ + e− → *OOH
path I O2 + H+ + e− → OOH
OOH + * → *OOH
path II O2 + * → *O2
*O−O + H+ + e− →
*O−O−H
*OOH + H+ + e− →
H2O2
O2 + 2H+ + e− →
H2O2
O2 + H+ + e− → *OOH
path I O2 + H+ + e− → OOH
OOH + * → *OOH
path II O2 + * → *O2
*O−O + H+ + e− →
*O−O−H
*OOH + H+ + e− → *O + H2O
*O + H2O + H+ + e− → *OH +
H2O
*OH + H2O + H+ + e− →
2H2O
O2 + 4H+ + 4e− → 2H2O
no
defects
defects
−0.17
−1.20
1.03
CF
−0.15
−1.20
1.05
DA
−1.33
−1.36
−1.50
−1.51
−1.21
−1.20
−0.01
−0.58
−0.63
−1.45
−1.20
−0.25
−0.88
−0.57
−0.80
−1.38
−0.83
−1.59
−1.60
−1.03
−4.99
−4.90
a
No defects and defects stand for these graphene clusters without and
with Stone−Wales defects, respectively; the asterisk refers to chemiadsorption on graphene; CF stands for calculation convergence failure;
DA stands for O2 dis-adsorption on graphene.
3549
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
Figure 5. Reaction energy diagram of ORR on (a) sulfur- and (b) sulfur oxide-doped graphene clusters.
first electron transmission of the two-electron transfer ORR on
the graphene cluster with sulfur atoms being the catalytic active
sites, Damjanovic’s proton participation reactions 1 and 2 seem
to be energetically unfavorable if these two reactions occur
separately because ΔG is negative (−1.20 eV) for reaction 1 but
positive (1.03−1.05 eV) for reaction 2. However, these two
reactions could occur if one-electron transfer takes place to
form OOH in the solvent, followed by the adsorption of the
neutral OOH on the graphene. For four-electron transfer which
takes place at carbon active sites, reactions 1 and 2 could occur
successively as ΔG for them is all negative. For the two-electron
transfer ORR, Yeager’s dissociation might not occur because O2
3550
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
0.27 eV) for the platinum (111) surface in Sha’s work.46 Our
results are comparable to the simulation results for nitrogendoped graphene (ΔEb = 0.19 eV), and nitrogen-doped carbon
nanotube (ΔEb = 0.30 eV).47 In the third electron transfer step,
there is no energy barrier found for both sulfur- and sulfur
oxide-doped graphene clusters, but in the last step, the
formation of water molecules, the reaction barriers are 0.38
and 0.05 eV for these sulfur- and sulfur oxide-doped graphenes,
respectively. These values are also comparable to that (ΔEb =
0.21 eV) for the same reaction over platinum (111).46 Overall,
the doping with −SO2− bonding structures seems better in
terms of the energy barriers and the number of electron
transfer. Since the energy barriers are comparable for the same
reactions over these materials, the sulfur-doped graphene may
show ORR catalytic properties similar to platinum, N-doped
graphene, and carbon nanotubes. This prediction has been
proved to be consistent with the experimental results on Sdoped graphene.12,32
could not adsorb on any potential catalytic active sites. For the
four-electron transfer, O2 adsorption and the following proton
reactions (reactions 3 and 4) could occur successively, as the
values of ΔG are all negative.
After the first electron transfer whether the reactions follow
Path I or II, the intermediate products are the same, OOH
adsorbed on the graphene. The reaction free energy of all
subreactions for two- or four-electron transfer paths is negative
for these sulfur-doped graphenes, indicating that the reaction
process would be energetically favorable. In our models, the
two-electron transfer usually occurs at the S-doped sites, which
possesses the highest positive charge density, while the fourelectron transfer proceeds mostly on carbon atoms with the
high positive spin or charge density. Thus, both sulfur- and
sulfur oxide-doped graphene would show high catalytic
activities. For the sulfur-adsorbed graphene, the presence of
Stone−Wales defects is critical for the graphene to catalyze
ORR.
We have compared our calculations with the experimental
results for sulfur-doped graphene clusters. For the two-electron
transfer reaction, O2 + 2H+ + e− → H2O2, our simulation
predicts that the free energies are −1.50 and −1.51 eV for those
sulfur-doped graphene clusters without and with Stone−Wales
defects, respectively, which are close to the experimental value
(ΔG = −1.40 eV) in standard states.35 For the four-electron
transfer pathway on the sulfur-doped graphene cluster without
and with defects, overall reaction O2 + 4H+ + 4e− → 2H2O,
calculated values of ΔG are −4.99 and −4.90 eV, which are also
close to the experimental results (ΔG = −4.92 eV) in standard
states.44 Thus, both two- and four-electron transfer processes
could simultaneously occur in thermodynamics on these sulfurdoped graphene clusters, and the number of electron transfer
could be between 2 and 4. This conclusion is consistent with
the experimental results that show the number of transferred
electrons ranging from 2.51 to 3.82 for the sulfur-doped
graphene.12
Although the above reactions over the S-doped graphene
clusters are thermodynamically favorable, the kinetics of the
ORR catalytic activities is determined by energy barriers ΔEb of
each reaction. It is necessary to determine the transition states
and reaction energy barriers of subreactions over the Sulfurdoped graphene cluster. Here taking the reactions on sulfur
(shown in Figure 5a) and oxide sulfur (shown in Figure 5b) at
the zigzag edge of graphene clusters as examples, we
determined the transition states and calculated the reaction
energy barriers. For the first reaction step, we found that there
is no reaction energy barrier for the OOH molecule adsorbing
on the catalytic active sites of the sulfur and oxide sulfur at the
zigzag edge of graphene clusters. Since the reaction energies of
the first step are −1.39 and −1.49 eV, respectively, for these
two doped graphene clusters, the first step is not a key step
affecting the reaction kinetics. The transition states in the
second step, the O−O bond breakage, are shown in Figure 5.
For the S-doped graphene cluster, the transition state is that the
*OOH still adsorbs on the active site of the graphene but a
proton is close to OOH with a O−H distance of 1.71 Å (inset
of Figure 5a). The energy barrier was calculated to be 0.1 eV for
the reaction where one water molecule and one adsorbed *O
are generated on the graphene. A similar transition structure
with a O−H distance of 1.56 Å (inset of Figure 5b) was found
for the oxide sulfur-doped graphene. The energy barrier is 0.24
eV for this reaction, much larger than that for the S-doped
graphene, but slightly smaller than the simulation value (ΔEb =
4. CONCLUSION
Four types of sulfur-doping structures, surface S-adsorbed, edge
S-substituted, edge SO2-substituted, and sulfur-ring connecting
graphene clusters, were proposed based on the experimental
results. The formation energy, electronic structures, as well as
ORR catalytic activities were calculated via DFT methods.
Among these doping structures, surface sulfur adsorption is the
most stable structure in terms of formation energy. The active
catalytic sites on these S-doped graphene clusters are those
carbon atoms located at the zigzag edges or close to the SO2
doping structure, which possess high positive charge density or
spin density. Both two-electron and four-electron transfers can
occur simultaneously over the S-doped graphene cluster. Twoelectron transfer pathways proceed on the substitutional sulfur
atom being the catalytic active sites with high charge density
while four-electron transfer takes place on the carbon atoms
with high positive spin or charge density. The Stone−Wales
defects facilitate the formation of surface S-adsorption on
graphene as well as the catalytic activities of sulfur-doped
graphene, especially for those with sulfur adsorbing on the
surface. The results for transition states and reaction energy
barriers of ORR subreactions reveal that the sulfur-doped
graphene clusters can show competitively catalytic properties
compared with platinum, nitrogen-doped carbon nanotube, and
graphene.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Tel: 940-369-5805. Fax: 940565-4824.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
The authors acknowledge the support from AFOSR MURI
(FA9550-12-1-0037) and the National Science Foundation
(IIP-1343270 and CMMI-1212259).
■
REFERENCES
(1) Xiong, W.; Du, F.; Liu, Y.; Perez, A.; Supp, M.; Ramakrishnan, T.
S.; Dai, L. M.; Jiang, L. 3-D Carbon Nanotube Structures Used as High
Performance Catalyst for Oxygen Reduction Reaction. J. Am. Chem.
Soc. 2010, 132, 15839−15841.
3551
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
(2) Snyder, J.; Fujita, T.; Chen, M. W.; Eelebancher, J. Oxygen
Reduction in Nanoporous Metal-Ionic Liquid Composite Electrocatalysts. Nat. Mater. 2010, 9, 904−907.
(3) Wang, C.; Daimon, H.; Onodera, T.; Kode, T.; Sun, S. H. A
General Approach to the Size- and Shape-Controlled Synthesis of
Platinum Nanoparticles and Their Catalytic Reduction of Oxygen.
Angew. Chem., Int. Ed. 2008, 47, 3588−3591.
(4) Lim, B.; Jiang, M. J. P.; Cho, E. C.; Tao, J.; Lu, X. M.; Zhu, Y. M.;
Xia, Y. N. Pd-Pt Bimetallic Nanodendrites with High Activity for
Oxygen Reduction. Science 2009, 324, 1302−1305.
(5) Zhang, L.; Zhang, J.; Wilkinson, D. P.; Wang, H. Progress in
Preparation of Non-noble Electrocatalysts for PEM Fuel Cell
Reactions. J. Power Sources 2006, 156, 171−182.
(6) Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P. HighPerformance Electrocatalysts for Oxygen Reduction Derived from
Polyaniline, Iron, and Cobalt. Science 2011, 332, 443−447.
(7) Lefevre, M.; Proietti, E.; Jaouen, F.; Dodelet, J. P. Iron-Based
Catalysts with Improved Oxygen Reduction Activity in Polymer
Electrolyte Fuel Cells. Science 2009, 324, 71−74.
(8) Byon, H. R.; Suntivich, J.; Shao-Horn, Y. Graphene-Based NonNoble-Metal Catalysts for Oxygen Reduction Reaction in Acid. Chem.
Mater. 2011, 23, 3421−3428.
(9) Gong, K. P.; Du, F.; Xia, Z. H.; Durstock, M.; Dai, L. M.
Nitrogen-Doped Carbon Nanotube Arrays with High Electrocatalytic
Activity for Oxygen Reduction. Science 2009, 323, 760−764.
(10) Wang, S.; Yu, D.; Dai, L. Polyelectrolyte Functionalized Carbon
Nanotubes as Efficient Metal-free Electrocatalysts for Oxygen
Reduction. J. Am. Chem. Soc. 2011, 133, 5182−5185.
(11) Yang, L.; Jiang, S.; Zhao, Y.; Zhu, L.; Chen, S.; Wang, X.; Wu,
Q.; Ma, J.; Ma, Y.; Hu, Z. Boron-Doped Carbon Nanotubes as MetalFree Electrocatalysts for the Oxygen Reduction Reaction. Angew.
Chem., Int. Ed. 2011, 50, 7132−7135.
(12) Yang, Z.; Yao, Z.; Li, G.; Fang, G.; Nie, H.; Liu, Z.; Zhou, X.;
Chen, X. a.; Huang, S. Sulfur-Doped Graphene as an Efficient Metalfree Cathode Catalyst for Oxygen Reduction. ACS Nano 2012, 6,
205−211.
(13) Li, X. L.; Zhang, G. Y.; Bai, X. D.; Sun, X. M.; Wang, X. R.;
Wang, E.; Dai, H. J. Highly Conducting Graphene Sheets and
Langmuir Blodgett Films. Nat. Nanotechnol. 2008, 3, 538−542.
(14) Lee, C.; Wei, X.; Kysar, J. W.; Hone, J. Measurement of the
Elastic Properties and Intrinsic Strength of Monolayer Graphene.
Science 2008, 321, 385−389.
(15) Patil, A. J.; Vickery, J. L.; Scott, T. B.; Mann, S. Aqueous
Stabilization and Self-Assembly of Graphene Sheets into Layered Bionanocomposites Using DNA. Adv. Mater. 2009, 21, 3159−3164.
(16) Lu, C. H.; Yang, H. H.; Zhu, C. L.; Chen, X.; Chen, G. N. A
Graphene Platform for Sensing Biomolecules. Angew. Chem., Int. Ed.
2009, 48, 4785−4787.
(17) Wu, Z. S.; Ren, W. C.; Xu, L.; Li, F.; Cheng, H. M. Doped
Graphene Sheets as Anode Materials with Superhigh Rate and Large
Capacity for Lithium Ion Batteries. ACS Nano 2011, 5, 5463−5471.
(18) Yoo, E.; Kim, J.; Hosono, E.; Zhou, H.; Kudo, T.; Honma, I.
Enhanced Cyclic Performance and Lithium Storage Capacity of SnO2/
Graphene Nanoporous Electrodes with Three-Dimensionally Delaminated Flexible Structure. Nano Lett. 2008, 8, 2277−2282.
(19) Qu, L. T.; Liu, Y.; Baek, J. B.; Dai, L. M. Nitrogen-Doped
Graphene as Efficient Metal-free Electrocatalyst for Oxygen Reduction
in Fuel Cells. ACS Nano 2010, 4, 1321−1326.
(20) Novoselov, K. S.; Jiang, Z.; Zhang, Y.; Morozov, S. V.; Stormer,
H. L.; Zeitler, U.; Maan, J. C.; Boebinger, G. S.; Kim, P.; Geim, A. K.
Room-Temperature Quantum Hall Effect in Graphene. Science 2007,
315, 1379.
(21) Loh, K. P.; Bao, Q. L.; Ang, P. K.; Yang, J. X. The Chemistry of
Graphene. J. Mater. Chem. 2010, 20, 2277−2289.
(22) Lee, C.; Wei, X. D.; Kysar, J. W.; Hone, J. Measurement of the
Elastic Properties and Intrinsic Strength of Monolayer Graphene.
Science 2008, 321, 385−388.
(23) Liu, H. T.; Liu, Y. Q.; Zhu, D. Chemical Doping of Graphene. J.
Mater. Chem. 2011, 21, 3335−3345.
(24) Yu, S. S.; Zhang, Q.; Dai, L. M. Highly Efficient Metal-free
Growth of Nitrogen-Doped Single-Walled Carbon Nanotubes on
Plasma-Etched Substrates for Oxygen Reduction. J. Am. Chem. Soc.
2010, 132, 15127−15129.
(25) Sheng, Z. H.; Tao, L.; Chen, J. J.; Bao, W. J.; Wang, F. B.; Xia, X.
H. Catalyst-free Synthesis of Nitrogen-Doped Graphene via Thermal
Annealing Graphite Oxide with Melamine and Its Excellent Electrocatalysis. ACS Nano 2011, 5, 4350−4358.
(26) Liu, R. L.; Wu, D. Q.; Feng, X. L.; Mullen, K. Nitrogen-Doped
Ordered Mesoporous Graphitic Arrays with High Electrocatalytic
Activity for Oxygen Reduction. Angew. Chem., Int. Ed. 2010, 49, 2565−
2569.
(27) Sheng, Z.; Gao, H.; Bao, W.; Wang, F.; Xia, X. Synthesis of
Boron Doped Graphene for Oxygen Reduction Reaction in Fuel Cells.
J. Mater. Chem. 2012, 22, 390−395.
(28) Wang, S.; Zhang, L.; Xia, Z.; Roy, A.; Chang, D.; Baek, J.; Dai, L.
BCN graphene as efficient metal-free electrocatalyst for the Oxygen
Reduction Reaction. Angew. Chem., Int. Ed. 2012, 51, 1−5.
(29) Zhang, L.; Xia, Z. Mechanisms of Oxygen Reduction Reaction
on Nitrogen-doped Graphene for Fuel Cell. J. Phys. Chem. C 2011,
115, 11170−11176.
(30) Zhang, L.; Niu, J.; Dai, L.; Xia, Z. Effect of Microstructure of
Nitrogen-doped Graphene on Oxygen Reduction Reaction Activity in
Fuel Cell. Langmuir 2012, 28, 7542−7550.
(31) Yang, S.; Zhi, L.; Tang, K.; Feng, X.; Maier, J.; Müllen, K.
Efficient Synthesis of Heteroatom (N or S)-Doped Graphene Based
on Ultrathin Graphene Oxide-porous Silica Sheets for Oxygen
Reduction Reactions. Adv. Funct. Mater. 2012, 22, 3634−3640.
(32) Jeon, I.; Zhang, S.; Zhang, L.; Choi, H.; Seo, J.; Xia, Z.; Dai, L.;
Baek, J. Edge-Selectively Sulfurized Graphene Nanoplatelets as
Efficient Metal-Free Electrocatalysts for Oxygen Reduction Reaction:
The Electron Spin Effect. Adv. Mater. 2013, 25, 6138−6145.
(33) Stephen, W.; Abhishek, D.; Masoud, A.; Heinz, P. Mechanism of
Molecular Oxygen Reduction at the Cathode of a PEM Fuel Cell:
Non-Electrochemical Reactions on Catalytic Pt Particles. J. Phys.
Chem. C 2008, 112, 8464−8475.
(34) Chen, R. R.; Li, H. X.; Chu, D.; Wang, G. F. Unraveling Oxygen
Reduction Mechanisms on Carbon-Supported Fe-Phthalocyanine and
Co-Phthalocyanine Catalysts in Alkaline Solutions. J. Phys. Chem. C
2009, 113, 20689−20697.
(35) Reyimjan, A.; Sidik; Alfred, B. A. O2 Reduction on Graphite and
Nitrogen-Doped Graphite: Experiment and Theory. J. Phys. Chem. B
2006, 110, 1787−1793.
(36) Ellen, V.; Alfred, B. A. Theoretical Predictions Concerning
Oxygen Reduction on Nitrided Graphite Edges and a Cobalt Center
Bonded to Them. J. Phys. Chem. C 2007, 111, 9330−9336.
(37) Kurak, K. A.; Anderson, A. B. Nitrogen-Treated Graphite and
Oxygen Electroreduction on Pyridinic Edge Sites. J. Phys. Chem. C
2009, 113, 6730−6734.
(38) Yu, L.; Pan, X.; Cao, X.; Hu, P.; Bao, X. Oxygen Reduction
Reaction Mechanism on Nitrogen-doped Graphene: A Density
Functional Theory Study. J. Catal. 2011, 282, 183−190.
(39) Kim, J. S.; Borisenko, K. B.; Nicolosi, V.; Kirkland, A. I.
Controlled Radiation Damage and Edge Structures in Boron Nitride
Membranes. ACS Nano 2011, 5, 3977−3986.
(40) Zhao, L.; He, R.; Gim, K.; Schiros, T.; Kim, K.; Zhou, H.;
Gutiérrez, C.; Chockalingam, S. P.; Arguello, C.; Pálová, L.; Nordlund,
D.; et al. Visualizing Individual Nitrogen Dopants in Monolayer
Graphene. Science 2011, 333, 999−1003.
(41) Feng, Y.; Li, F.; Hu, Z.; Luo, X.; Zhang, L.; Zhou, X.; Wang, H.;
Xu, J.; Wang, E. Tuning the Catalytic Property of Nitrogen-doped
Graphene for Cathode Oxygen Reduction Reaction. Phys. Rev. B 2012,
85, 155454.
(42) Wu, Y.; Fang, S.; Jiang, Y.; Holze, R. Effects of Doped Sulfur on
Electrochemical Performance of Carbon Anode. J. Power Sources 2002,
108, 245−245.
(43) Gao, H.; Liu, Z.; Song, L.; Guo, W.; Gao, W.; Ci, L.; Rao, A.;
Quan, W.; Vajtai, R.; Ajayan, P. M. Synthesis of S-doped Graphene by
Liquid Precursor. Nanotechnology 2012, 23, 275605.
3552
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553
The Journal of Physical Chemistry C
Article
(44) No̷ rskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L. Origin
of the Overpotential for Oxygen Reduction at a Fuel-Cell Cathode. J.
Phys. Chem. B 2004, 108, 17886−17892.
(45) Ochtershi, J. W. Thermochemistry in Gaussian; Gaussian, Inc.:
Wallingford, CT, 2000.
(46) Sha, Y.; Yu, T.; Liu, Y.; Merinov, B.; Goddard, W., III
Theoretical Study of Solvent Effects on the Platinum-Catalyzed
Oxygen Reduction Reaction. J. Phys. Chem. Lett. 2010, 1, 856−861.
(47) Ni, S.; Li, Z.; Yang, J. Oxygen Molecule Dissociation on Carbon
Nanostructures with Different Types of Nitrogen Doing. Nanoscale
2012, 4, 1184−1189.
3553
dx.doi.org/10.1021/jp410501u | J. Phys. Chem. C 2014, 118, 3545−3553